.______ _______ .__ __. _______ __ .___________.___________. _______ .______
| _ \ | ____|| \ | | / _____|| | | | || ____|| _ \
| |_) | | |__ | \| | | | __ | | `---| |----`---| |----`| |__ | |_) |
| _ < | __| | . ` | | | |_ | | | | | | | | __| | /
| |_) | | |____ | |\ | | |__| | | | | | | | | |____ | |\ \----.
|______/ |_______||__| \__| \______| |__| |__| |__| |_______|| _| `._____|
Data science manager focused on open-source LLM research, particularly tool integration patterns and reasoning workflows. Currently exploring model internals and methods for behavior modification without traditional fine-tuning approaches.
Believer in digital sovereignty and secure, independent infrastructure. My work is supported by a private compute environment: AMD EPYC cluster with 4x 3090s, 100TB NAS, Proxmox virtualization, and 10Gb networking. This setup enables focused research while maintaining control over data and compute resources.
Contributing to the open source community through projects like OpenAI-compliant tools for Tabby API, dataset processing utilities, and XGrammar/SGLang integration work. Maintaining a collection of quantized models on HuggingFace with a focus on practical tool-calling capabilities.
Past ventures in data consultation, manufacturing, and small business have shaped my approach to building sustainable systems. Each experience has reinforced the value of ownership, security, and iterative improvement.
Interested in connecting with researchers and developers working on LLM capabilities, particularly those focused on secure and independent development workflows. If you're exploring similar areas or want to discuss collaborative opportunities:
→
GitHub - Open source contributions and projects
→
HuggingFace - Model releases and experiments
→
LinkedIn - Professional background and contact
→ For folks with access, here is the chat portal -
Chat.Gitter.Ai